OPINION OF THE COURT
McKEE, Circuit Judge.MCI Telecommunications (“MCI”), a long distance telecommunications service provider, has sued Teleconeepts to recover the cost of services MCI provided under MCI’s Federal Communications Commission tariff (“FCC tariff’). Teleconeepts raised the untimely service of the complaint and the statute of limitations as defenses, and also brought a third-party action against Bell of Pennsylvania (“Bell”), Teleconeepts’ local telephone exchange carrier. Bell disclaimed liability under the terms of its Pennsylvania Public Utility Commission Tariff (“PUC Tariff’). The district court held that the action was not time-barred and that “good cause” existed for the late service of the complaint. The court also granted summary judgment to both MCI and Bell holding that the FCC and PUC tariffs both placed the responsibility for unauthorized telephone calls on Teleconeepts.
Because we find that MCI’s action was partially barred by the statute of limitations, and that the doctrine of primary jurisdiction required the district court to transfer the third-party complaint to the Pennsylvania Public Utility Commission, we will reverse in part and remand for further proceedings.
I. FACTUAL BACKGROUND
Teleconeepts owns coin operated telephones — commonly referred to as “pay phones” — that it places on the premises of various businesses. MCI supplied long dis*1091tance telephone service to Teleeoncepts from January 1988 through March 1990 under the terms and conditions of the tariff MCI had filed with the Federal Communications Commission. When Teleeoncepts’ pay phones are used, Teleeoncepts incurs a cost to Bell of Pennsylvania for the use of Bell’s telephone lines. The monthly bill for the line charges also includes the customer’s long distance charges for the preceding month. Teleeon-cepts’ November 1989 bills from MCI for long distance calls included long distance service charges for international telephone calls in excess of $7,000. Teleeoncepts was billed under six different account numbers, which presumably represent six different Teleeon-cepts’ pay phones. The November charges exceeded prior months’ long distance charges to such an extent that Teleeoncepts was certain that a billing error had occurred, and so informed Bell. However, since Bell was merely a conduit for billing long distance charges, it responded by telling Teleeoncepts to contact its long distance carrier — MCI.
Teleeoncepts contacted MCI and informed it of the numerous long distance calls to the Dominican Republic and Puerto Rico that Teleeoncepts believed had not been made from any of its phones and requested a credit. When MCI refused, Teleeoncepts told MCI that it would not pay for these long distance charges, but MCI continued to provide long distance service. When Teleeon-cepts received its December bills it discovered over $13,000 in doubtful charges to Puerto Rico and the Dominican Republic. Teleeon-cepts again refused to pay these charges.
On December 27, 1989, MCI notified Tele-concepts that its long distance service was terminated. However, for some reason, MCI failed to terminate long distance service until the following March. In the interim, Tele-concepts continued to receive bills containing exorbitant long distance charges, and Tele-concepts continued to refuse to pay. Finally, MCI sued Teleeoncepts to recover the amount of unpaid charges for long distance services MCI had provided to Teleeoncepts through March 1990 — $47,565.84.
Eventually, Teleeoncepts came to believe that the questioned telephone calls had resulted from a fraudulent process known as “hacking.”1 This occurred when a person called an 800 number on a pay phone and remained silent until the receiving party hung up. A second dial tone would then be given to the 800 caller who could then call anywhere he or she desired without placing any additional coins in the telephone.
On January 15, 1992, MCI filed its initial summons and complaint in an effort to collect the unpaid charges from Teleeoncepts. MCI attempted service through the Mercer County Sheriffs Department, but its initial attempt was unsuccessful. Service was eventually made on June 25, 1992. Teleeoncepts responded by filing a third-party complaint against Bell of Pennsylvania in which it alleged that Bell was responsible for the defect in the dial tone that allowed the illegal “hacking” and that Bell should therefore indemnify Teleeoncepts for any liability it may have to MCI.2 Teleeoncepts eventually moved to dismiss the complaint because MCI had failed to effect service of process within 120 days of filing of the complaint as required by Federal Rule of Civil Procedure 4(j). In an order dated September 15, 1992, the district court denied Teleeoncepts’ motion to dismiss MCI’s complaint finding that “good cause” excused the late service.
Subsequently, the parties filed cross-motions for summary judgment. Teleeoncepts claimed that MCI’s action was untimely since it was not filed within the two year statute of limitations contained in the Communications Act. Teleeoncepts argued that MCI’s cause of action accrued either when it refused to pay the November 1989 bills, or at the latest, on December 27, 1989, when MCI gave notice that Teleeoncepts’ long distance services were terminated. MCI countered by argu*1092ing that its action was timely because Tele-concepts’ services continued until March 1990 despite the December 27, 1989 disconnect notice. MCI further argued that under a 30 day payment provision of its federal tariff, final payment of the bills would not become due until either April 1990, or January 27, 1990, at the earliest even accepting Telecon-cepts’ position. Thus, MCI claimed the operative date for commencing an action was either March or April of 1992, or at the earliest, January 27, 1992.
The district court denied Teleconeepts’ motion for summary judgment in a memorandum opinion and order dated December 28, 1993. Additionally, the court held that MCI’s federal tariff placed responsibility for unauthorized calls on Teleconeepts, and thus, granted MCI’s cross-motion for summary judgment. The court also granted MCI’s request for attorney’s fees and, in a separate memorandum opinion and order dated February 25, 1994, determined the reasonable amount of such fees to be $11,812.50. The court also held that the PUC tariff placed responsibility for unauthorized calls on payphone owners, and therefore granted summary judgment in favor of Bell and against Teleconeepts in a memorandum and order entered on June 20, 1994.
On appeal, Teleconeepts challenges the district court’s denial of its motion to dismiss for failure to timely serve the complaint, the denial of its motion for summary judgment on the statute of limitations defense, the grant of summary judgment in favor of Bell, and the amount of the attorney’s fee award.
II. DISCUSSION
A. Appellate Jurisdiction
We must first determine whether we have jurisdiction to review the issues raised by Teleconeepts on appeal. The notice of appeal reads as follows:
Teleconeepts, Inc., defendant-third party plaintiff appeals to the United States Court of Appeals for the Third Circuit from an order of summary judgment disposing the remaining claims of the District Court for the District of New Jersey entered in this case June 20,1994, in favor of third party defendant, Bell of Pennsylvania and December 28,1993 in favor of plaintiff, MCI Telecommunications, Inc.
App. at 1.
Federal Rule of Appellate Procedure 3(c) provides, in pertinent part, that a notice of appeal “must designate the judgment, order or part thereof appealed from....” Fed.R.App.P. 3(c). If a party does not satisfy the requirements of Federal Rule of Appellate Procedure 3(c), then the appellate court does not acquire jurisdiction over the undesignated issues. United States v. Rivera Constr. Co., 863 F.2d 293, 298 (3d Cir.1988). Even though the notice of appeal does not mention the September 15, 1992 order (denying Teleconeepts’ motion to dismiss) or the district court’s February 25, 1994 memorandum and order (calculating MCI’s award of reasonable attorney’s fees), Teleconeepts challenges both of these decisions in its brief to this court. MCI argues that we did not acquire jurisdiction over these issues since these orders are neither directly nor indirectly referred to in the notice of appeal. Appellee’s brief at 6.
“Our jurisprudence liberally construes notices of appeals.” Drinkwater v. Union Carbide Corp., 904 F.2d 853, 858 (3d Cir.1990). Thus, we have held that it is proper to exercise appellate jurisdiction over orders not specified in the notice of appeal if “ ‘there is a connection between the specified and unspecified order, the intention to appeal the unspecified order is apparent and the opposing party is not prejudiced and has a full opportunity to brief the issues.’ ” Lusardi v. Xerox Corp., 975 F.2d 964, 972 (3d Cir.1992) (quoting Williams v. Guzzardi, 875 F.2d 46, 49 (3d Cir.1989)). These factors are present here.
We have repeatedly held that “ ‘since ... only a final judgment or order is appealable, the appeal of a final judgment draws into question all prior non-final orders and rulings.’ ” Drinkwater, 904 F.2d at 858 (quoting Elfman Motors, Inc. v. Chrysler Corp., 567 F.2d 1252, 1253 (3d Cir.1977)). Telecon-cepts could not appeal the September 15, 1992 order denying its motion to dismiss MCI’s complaint until the district court filed *1093the December 28, 1993 order of summary judgment in favor of MCI. Moreover, in disposing of the remaining issues, the December 28, 1998 memorandum opinion refers to the district court’s September 15, 1992 order denying Teleconcepts’ motion to dismiss. Thus, the requisite connection is present.
Similarly, while the notice of appeal does not refer to the February 25, 1994 memorandum and order calculating MCI’s award of reasonable attorney’s fees, an adequate connection exists between a specified order that designates the prevailing party for purposes of attorney’s fees and an unspecified order that quantifies the attorney’s fee award. See Bernardsville Bd. of Educ. v. J.H., 42 F.3d 149, 156 n. 10 (3d Cir.1994). Since the December 28,1993 order specifically granted MCI’s request for attorney’s fees and merely directed MCI to file an affidavit of reasonable attorney’s fees, there is an adequate connection between these two orders.
Moreover, MCI is not prejudiced as it had an opportunity to brief the disputed issues, and has done so. See id. at 156 n. 10. Accordingly, we hold that we have jurisdiction to review the September 15, 1992 order denying Teleconcepts’ motion to dismiss and the February 25, 1994 memorandum and order calculating an award of attorney’s fees.
B. Subject Matter Jurisdiction
Before addressing the substantive issues raised by Teleconcepts we must determine if the district court had subject matter jurisdiction over MCI’s action against Tele-concepts in the first place. While neither the district court nor Teleconcepts ever raised this issue we have an obligation to do so sua sponte. See Medlin v. Boeing Vertol Co., 620 F.2d 957, 960 (3d Cir.1980); Carlsberg Resources Corp. v. Cambria Sav. & Loan Ass’n, 554 F.2d 1254, 1256 (3d Cir.1977).
MCI’s action is based upon Teleconcepts’ failure to pay MCI for long distance telephone service MCI provided under the terms and conditions set forth in MCI’s FCC Tariff. MCI alleges that since it is required to collect the charges on the services specified in the tariff under § 203 of the Communications Act of 1934, 47 U.S.C. § 203 (1982), subject matter jurisdiction exists under 28 U.S.C. §§ 1331 and 1337, and the Communications Act of 1934, 47 U.S.C. § 151, et seq. (1982).3
While this circuit has never addressed whether the collection of unpaid charges for long distance telephone service under an FCC tariff “arises under” federal law (28 U.S.C. § 1331) or an act of Congress regulating commerce (28 U.S.C. § 1337) the majority of courts of appeals that have addressed this issue have answered in the affirmative. See Western Union Int'l Inc. v. Data Dev., Inc., 41 F.3d 1494 (11th Cir.1995); MCI Telecommunications Corp. v. Graham, 7 F.3d 477 (6th Cir.1993); MCI Telecommunications Corp. v. Garden State Inv. Corp., 981 F.2d 385 (8th Cir.1992); Ivy Broadcasting Co., Inc. v. Am. Tel. & Tel. Co., 391 F.2d 486 (2d Cir.1968). But see MCI Telecommunications Corp. v. Credit Builders of Am., Inc., 980 F.2d 1021 (5th Cir.1993), vacated, — U.S. -, 113 S.Ct. 2925, 124 L.Ed.2d 676 (1993), prior opinion reinstated, 2 F.3d 103 (5th Cir.), cert. denied, — U.S. -, 114 S.Ct. 472, 126 L.Ed.2d 424 (1993).
In Rickman Bros. Records, Inc. v. U.S. Sprint Communications Co., 953 F.2d 1431, 1438 (3d Cir.1991), we held that the district court had jurisdiction over Sprint’s action under a tariff filed with the FCC pursuant to 28 U.S.C. § 1337(a). However, we did not question jurisdiction because there was diversity of citizenship and we would thus have had jurisdiction even if federal question jurisdiction failed. See id. Given the subsequent decisions of our sister circuits and the lack of diversity here, we now undertake this analysis.
Those courts of appeals that have held that subject matter jurisdiction exists under 28 U.S.C. §§ 1331 and/or 1337(a), reason that a *1094claim for unpaid long distance charges “arises under” an act of Congress regulating commerce (the Communications Act) because the claim relies on tariffs that must be filed with the FCC.4 See Western Union, 41 F.3d at 1496; Graham, 7 F.3d at 479; Garden State Inv., 981 F.2d at 388; Ivy Broadcasting, 391 F.2d at 493-494. The lone appellate court to hold otherwise reasons that the federal common law is appropriate in only a “few and restricted” circumstances and doubts that collection of a delinquent phone bill falls within these limited circumstances. See Credit Builders, 980 F.2d at 1022-23.
This specific issue was first addressed in Ivy Broadcasting Co., Inc. v. Am. Tel. & Tel., supra. There, the district court sua sponte dismissed a suit against A.T. & T. that arose from problems that had occurred during a transmission over A.T. & T.’s telephone wires. The court reasoned that claims for negligence and breach of contract did not arise out of the Communications Act but were founded on tort and contract law, and that the counterclaim was merely an action for services rendered that also lacked a federal jurisdictional basis. On appeal the United States Court of Appeals for the Second Circuit noted that since the complaint did not allege a specific violation of the Communications Act, 47 U.S.C. § 207 did not confer jurisdiction. Nevertheless, the court reasoned that the broad scheme of federal regulation of communications carriers indicated a congressional intent to occupy the field to the exclusion of state law. See Ivy Broadcasting, 391 F.2d at 490.
It seems to us that the congressional purpose can be achieved only if a uniform federal law governs as to the standards of service which the carrier must provide and as to the extent of liability for failure to comply with such standards.
Id. at 491.
The court concluded that since federal law controlled, the suit “arose under” the laws of the United States as required by 28 U.S.C. § 1331.
The word ‘laws’ in § 1331 should be construed to include laws created by federal judicial decision as well as by congressional legislation. The rationale of the 1875 grant of federal question jurisdiction — to insure the availability of a forum designed to minimize the danger of hostility toward, and specially suited to the vindication of, federally created rights — is as applicable to judicially created rights as to rights created by statute.
Id. at 492.
The court also held that 28 U.S.C. § 1337 provided jurisdiction over the counterclaims.
The Communications Act of 1934 is an ‘Act of Congress regulating commerce’ within the meaning of [§ 1337.] Since we conclude that the counterclaims arise under the Communications Act insofar as they rely upon tariffs which the Act requires to be filed with the FCC, we hold that [§ 1337] gives the district court jurisdiction over so much of the counterclaims as relies upon such tariffs.
Id. at 494 (footnote omitted) (citations omitted). The court’s analysis relied heavily upon an analogous inquiry of the Supreme Court under the Commerce Act in Louisville & N.R. v. Rice, 247 U.S. 201, 202, 38 S.Ct. 429, 429, 62 L.Ed. 1071 (1918).
Several other courts have employed the reasoning of Ivy Broadcasting to hold that federal district courts have subject matter jurisdiction over actions for unpaid charges for services provided under an FCC tariff. In MCI Telecommunications Corp. v. Garden State Inv. Corp, supra, MCI sued to recover unpaid telecommunications charges, and the district court dismissed the complaint sua sponte because the complaint did not allege a specific violation of the Communications Act. The court reasoned that there was no need for uniform federal common law governing claims to collect unpaid telecom*1095munication service charges, and that there was therefore no basis for the exercise of subject matter jurisdiction.
The United States Court of Appeals for the Eighth Circuit relied heavily upon Supreme Court decisions involving federal jurisdiction under the Commerce Act to reverse. (i.e. Thurston Motor Lines, Inc. v. Jordan K. Rand, Ltd., 460 U.S. 533, 103 S.Ct. 1343, 75 L.Ed.2d 260 (1983) (per curiam)).5 See Garden State Inv., 981 F.2d at 387-88. The court reasoned that:
the district court failed to recognize that a claim arises under federal law when a right created by federal law is an essential element of the plaintiff’s action. The district court stated MCI’s claim brought under an FCC tariff ‘is simply a contract action seeking to recover payment for services rendered.’ The district court’s characterization of MCI’s claim overlooks the fact that federal tariffs are the law, not mere contracts. Although a user’s refusal to pay charges fixed by a tariff will often arise in the context of a broken contract, the carrier’s claim for payment is necessarily based on the filed tariff.
Id. at 387 (citations omitted).
In analogous circumstances, the United States Court of Appeals for the Sixth Circuit reversed the district court’s sua sponte dismissal for lack of federal jurisdiction in MCI Telecommunications Corp. v. Graham, supra. The court held that MCI’s ability to sue was based upon its FCC tariff, and therefore was rooted in federal law. See Graham, 7 F.3d at 479-80. As in Ivy Broadcasting, the court was persuaded by similarities between the Communications Act and the Commerce Act and was guided by Thur-ston, Rice and the analysis of the United States Court of Appeals for the Second Circuit in Ivy Broadcasting. “[I]t would be incongruous to impose a different jurisdictional rule under the Communications Act than under the Commerce Act.” Id. at 480. See also Western Union, 41 F.3d at 1496-97 (subject matter jurisdiction over a claim for unpaid telecommunications service charges existed under 28 U.S.C. § 1331 and 28 U.S.C. § 1337).
Only MCI Telecommunications Corp. v. Credit Builders of Am., Inc., supra, reached a contrary result. There, a supplier of telecommunications services appealed the district court’s determination that a suit to collect unpaid telecommunications charges lacked a jurisdictional basis. On appeal, the plaintiff argued that the district court had independent federal question jurisdiction under 28 U.S.C. § 1331 and jurisdiction over matters arising out of the Communications Act under 28 U.S.C. § 1337.
The United States Court of Appeals for the Fifth Circuit began its analysis by noting that:
the Supreme Court has stated that a case arises under federal law if ‘it really and substantially involves a dispute or controversy respecting the validity, construction, or effect of such a law, upon the determination of which the result depends.’
Credit Builders, 980 F.2d at 1022 (citations omitted).
The court was not persuaded by the reasoning of Ivy Broadcasting. Instead, the court reasoned that:
[a]s the Supreme Court has emphasized, the federal common law is appropriate in only a ‘few and restricted’ circumstances. Milwaukee v. Illinois, 451 U.S. 304, 313, 101 S.Ct. 1784, 1790, 68 L.Ed.2d 114 (1981). In Texas Industries, Inc. v. Radcliff Materials, Inc., 451 U.S. 630, 641, 101 S.Ct. 2061, 2067, 68 L.Ed.2d 500 (1981), the Supreme Court went on to state that ‘absent some congressional authorization to formulate substantive rules of decision, federal common law exists only in such narrow areas as those concerned with the rights and obligations of the United States, interstate and international disputes implicating the conflicting rights of States or our relations with foreign nations, and admiralty cases.’ We do not believe that this case to collect a delinquent telephone bill falls within these limited instances.
*1096Id. at 1022-23. The court held that for these same reasons § 1337 did not confer federal question jurisdiction, and that plaintiffs action for breach of contract or quantum meru-it was a creature of state law.
We are not persuaded by the analysis in Credit Builders. MCI’s action is based upon, and draws its life from, the tariff that MCI filed with the Federal Communications Commission. The reasoning of Ivy Broadcasting, and the analogous cases decided under the Commerce Act, see Thurston and Rice, supra, persuade us that the district court did have subject matter jurisdiction over MCI’s action, and Teleconcepts’ counterclaim. However, there are other jurisdictional problems with that counterclaim which we discuss in more detail below.
C. Late Service of the Complaint
When the district court denied Telecon-cepts’ motion to dismiss for failure to timely serve the complaint, Federal Rule of Civil Procedure 4(j) read in pertinent part:
Summons: Time Limit for Service. If a service of the summons and complaint is not made upon a defendant within 120 days after the filing of the complaint and the party on whose behalf such service was required cannot show good cause why such service was not made within that period, the action shall be dismissed as to that defendant without prejudice upon the court’s own initiative with notice to such party or upon motion.
Fed.R.Civ.P. 4(j).6
The district court was thus required to dismiss MCI’s action if process was not served within 120 days of the filing of the complaint unless MCI could show good cause for the delinquency. See Petrucelli v. Bohringer & Ratzinger, 46 F.3d 1298, 1304 (3d Cir.1995).
MCI filed its initial summons and complaint on January 15,1992. The papers were returned unserved by the Mercer County Sheriffs Department marked “unable to locate, unknown at address given” on February 25, 1992. MCI requested an alias summons on or about March 12, 1992, after it discovered another address for service. The alias summons was returned on an “unknown date” and forwarded for service on or about May 29, 1992. Service of process was eventually achieved at this alternate address on June 25, 1992, well over a month after the 120 days prescribed by Rule 4(j) had lapsed. MCI never made a request for an extension of time.
The district court found that “good cause” excused the late service, and denied Teleconcepts’ motion to dismiss the complaint. Our review of the district court’s finding of “good cause” is for an abuse of discretion. See Lovelace v. Acme Markets, Inc., 820 F.2d 81, 83 (3d Cir.), cert. denied, 484 U.S. 965, 108 S.Ct. 455, 98 L.Ed.2d 395 (1987); Braxton v. United States, 817 F.2d 238, 242 (3d Cir.1987).
The district court did not articulate the factor(s) it believed constituted “good cause.” The court initially orally denied Teleconcepts’ motion to dismiss during the following exchange in a telephone conference:
THE COURT: All right. Now, the defendant moved to dismiss the complaint pursuant to Rule 4(j), which provides for dismissal unless good cause be shown. I would like to hear from the defendant before I rule.
MR. REILLY: Your Honor, after I received the response from the plaintiff regarding my client’s address as 51 Everett Street in Princeton, he informed me that that’s been his address. I understand that he may not have been there at the time when the Sheriff initially went out. The problem I have with the argument is that they did serve him eventually at another— at his residence.
THE COURT: Yes.
MR. REILLY: And that was some four months after the initial issuance of the summons and complaint, which I still feel is an inordinate amount of time—
*1097THE COURT: Well, I’ve seen no prejudice. The remedy if good cause were not shown would be dismissal without prejudice and re-service. Under the circumstances, I certainly find good cause has been shown. Motion is denied. The defendant will answer, move or otherwise plead.
Subsequently, the district court memorialized this decision in the written order of September 15,1992. In that order, the court stated merely that “good cause” had been shown and that the motion to dismiss was denied for the reasons set forth on the record.
Although the district court felt that Teleconcepts had not been prejudiced by the late service, absence of prejudice alone can never constitute good cause to excuse late service. See United States v. Nuttall, 122 F.R.D. 163, 166-67 (D.Del.1988) (courts have considered three factors in determining the existence of good cause: (1) reasonableness of plaintiffs efforts to serve (2) prejudice to the defendant by lack of timely service and (3) whether plaintiff moved for an enlargement of time to serve). We have equated “good cause” with the concept of “excusable neglect” of Federal Rule of Civil Procedure 6(b)(2), which requires “a demonstration of good faith on the part of the party seeking an enlargement and some reasonable basis for noncompliance within the time specified in the rules.” See Petrucelli, 46 F.3d at 1312 (Becker, J., concurring in part and dissenting in part).7 Thus, while the prejudice may tip the “good cause” scale, the primary focus is on the plaintiffs reasons for not complying with the time limit in the first place. Such “justifications” are conspicuously absent in the district court’s oral decision and its subsequent written order. Moreover, the briefs to this court are silent on this issue and the parties have therefore not assisted in divining the “good cause” that the district court found. In addition, our review of the entire record has uncovered only one reference to the “good cause” which the court may have felt supported late service. In MCI’s brief in opposition to Teleconcepts’ motion to dismiss MCI states:
good cause is shown because service could not be made at the address given as the registered address for service of process at the time the complaint was filed. It was necessary to make additional attempts at service by locating another address and requesting an alias summons.
App. at 35.
Even if we were to speculate and conclude that this was the basis for the district court’s finding of “good cause,” we would have to conclude that it was an abuse of discretion. The summons was returned unserved on February 28, 1992. MCI learned of Teleconcepts’ alternative address as early as March 12,1992, and requested an alias summons on or about that same date. Inexplicably, the summons was not forwarded for service until on or about May 29,1992, and not served at this alternate address until June 25, 1992, well beyond the time limit prescribed by Rule 4(j). MCI never moved for an extension of time.
Nothing on this record explains why it took MCI over three months after it learned of Teleconcepts’ alternate address to serve Teleconcepts. Moreover, the record does not explain why MCI never filed a motion to enlarge the time to serve. See Lovelace, 820 F.2d at 85 (alternative means of service and the ability to extend the time indicate a lack of diligence and weigh against a finding of good cause). Since we are presented with no explanations as to what, if any, circumstances constitute sufficient “good cause” to excuse MCI’s apparent lack of diligence, we hold that the district court abused its discretion in finding that good cause existed to excuse the late service. See Braxton, 817 F.2d at 242 (good cause does not exist when there is an *1098“unexplained delinquency on the part of the process server and lack of oversight by counsel”).
Reversal of a district court’s finding that good cause existed to excuse late service results in the dismissal of an action, but such dismissal is without prejudice to the plaintiff. Accordingly, the party can refile the complaint and receive a new 120 day period to serve process. See Petrucelli, 46 F.3d at 1304 n. 6.
However, as of December 1,1993, Rule 4(j) was amended and redesignated Rule 4(m). Rule 4(m) provides, in part, that:
If service of the summons and complaint is not made upon a defendant within 120 days after the filing of the complaint, the court, upon motion, or on its own initiative after notice to the plaintiff, shall dismiss the action without prejudice as to that defendant or direct that service be effectuated within a specified time; provided that if the plaintiff shows good cause for the failure, the court shall extend the time for service for an appropriate period.
Fed.R.Civ.P. 4(m).
We recently addressed the significance of this amendment in Petrucelli v. Bohringer, supra. There, we read Rule 4(m) “to require a court to extend time if good cause is shown and to allow a court discretion to dismiss or extend time absent a showing of good cause.” Petrucelli, 46 F.3d at 1305. Here the statute of limitations is in issue. In Petrucelli, we emphasized that the expiration of the statute of limitations does not require the court to extend the time for service, as the court has discretion to dismiss the case even if the refiling of the action is barred. See id. at 1306. We also noted that Rule 4(m) should apply retroactively, to all matters pending at the time it became effective “insofar as just and practicable.” Id. at Í305 (quoting The Order of the United States Supreme Court Adopting and Amending the Federal Rules of Civil Procedure (April 22, 1993)). Here, such retroactive application is both just and practicable as the district court has already made a determination that there is some basis to excuse MCI’s lack of diligence. As a result, the district court would have discretion to allow MCI’s action to proceed upon remand under Rule 4(m) even though we have determined that no “good cause” was shown under Rule 4(j).
[A]s a result of the rule change which led to Rule 4(m), when entertaining a motion to extend time for service, the district court must proceed in the following manner. First, the district court should determine whether good cause exists for an extension of time. If good cause is present, the district court must extend time for service and the inquiry is ended. If, however, good cause does not exist, the court may in its discretion decide whether to dismiss the ease without prejudice or extend time for service.
Id. at 1305. Moreover, the expiration of the statute of limitations does not prohibit the district court from extending the time for service. See id. at 1305-06.
The parties here apparently do not consider the substantive changes to Rule 4(j) significant to our analysis as neither has mentioned the amendment of Rule 4(j) or cited Petrucelli. However, we find Petrucelli’s interpretation of the amendment to Rule 4(j) and its retroactive impact dispositive to the issues before us.
Accordingly, even though we have determined that the district court abused its discretion in inexplicably finding “good cause” for MCI’s lack of diligence, the retroactive effect of Rule 4(m) means that the district court had the discretion to allow this action to proceed even in the absence of “good cause.” We view the district court’s decision to extend time as an exercise of its discretion under that Rule, and therefore, we affirm the district court’s denial of Teleeoncepts’ motion to dismiss.
D. Statute of Limitations
Teleeoncepts argues that the district court improperly denied its motion for summary judgment based upon the statute of limitations. The Communications Act of 1934, provides that “[a]ll actions at law by carriers for recovery of their lawful charges, or any part thereof, shall be begun, within two years from the time the cause of action accrues, and not after.” 47 U.S.C. § 415(a) *1099(1982). The Act further states that a cause of action “in respect of the transmission of a message shall, for the purposes of this section, be deemed to accrue upon delivery ... thereof by the carrier and not after.” 47 U.S.C. § 415(e) (1982). However, as noted above, MCI’s FCC tariff provides that “MCI’s bills are payable upon receipt. Amounts not paid within 30 days after the date of the invoice will be considered past due....”
In the district court, Teleeoncepts argued that payment was due when it received the bills for long distance service in November 1989, and MCI’s cause of action accrued when these bills went unpaid. Teleeoncepts also argued that, even if the “past due” standard were used, the cause of action accrued in December 1989; and that the latest possible accrual date was December 27, 1989, when MCI issued its letter terminating service. Since MCI’s action was filed more than two years from any of these potential accrual dates, Teleeoncepts concludes that the action was untimely.
The district court rejected this position reasoning that “[i]n an action involving collection of accounts receivable, common sense dictates that in determining the tolling date of the applicable statute of limitations, this court must look to the date upon which payment was demanded and refused.” MCI Telecommunications Corp. v. Teleeoncepts, Inc., No. 92-244, slip op. at 5 (D.N.J. December 28,1993). The court concluded that MCI had until April 16, 1992—two years after Teleeoncepts failed to remit payment of its final invoice of March 17, 1990—to initiate this action. Moreover, the court suggested that even if it were to focus on the December 27, 1989 termination letter, payment would not be due until January 26, 1990 (allowing for the 30 day period contained in the tariff). Since MCI filed the complaint on January 15, 1992, the court concluded it was timely and denied Teleeoncepts’ motion for summary judgment.
Our review of the district court’s denial of Teleeoncepts’ summary judgment motion is plenary. See Gulfstream III Assoc., Inc. v. Gulfstream Aerospace Corp., 995 F.2d 425, 429 (3d Cir.1993); Schafer v. Bd. of Pub. Educ. of Sch. Dist. of Pittsburgh, Pa., 903 F.2d 243, 246 (3d Cir.1990). It is clear that our analysis is controlled by 47 U.S.C. § 415(a). However, we must determine when a cause of action “accrues” for purposes of that statue.
In resolving this issue of first impression, Teleeoncepts urges us to be guided by the numerous decisions that have construed the statute of limitations in the Commerce Act, see 49 U.S.C. § 16(3) (repealed 1978),8 on which § 415 was based. See Pennsylvania R. Co. v. Carolina Portland Cement Co., 16 F.2d 760 (4th Cir.1927); South Omaha Terminal Ry. Co., Inc. v. Armour & Co., Inc., 373 F.Supp. 641, (D.Neb.1974); Baker v. Chamberlain Mfg. Corp., 356 F.Supp. 1314, (N.D.Ill.1973). In those cases the courts ruled that a carrier’s action to recover charges accrues when delivery is made or tendered, irrespective of what occurs subsequent to delivery. Teleeoncepts reasons by analogy that MCI’s cause of action accrued when it transmitted the long distance signals (i.e. “tendered delivery”) in October 1989. Appellant’s brief at 24.
We do not find this argument persuasive. At the time the cases Teleeoncepts relies upon were decided the Commerce Act provided that “[t]he cause of action in respect of a shipment of property shall, for purposes of this section, be deemed to accrue upon delivery or tender of delivery thereof by the carrier, and not after.” 49 U.S.C. § 16(3)(e) (repealed 1978).9 However, the Commerce Act is designed:
to fix one date on which all causes of action, both those in favor of shipper and those in favor of carrier, with respect to any particular shipment, should be deemed to have accrued, so that, in the application of the section limiting time for suit, a *1100situation would not arise wherein claims in favor of one party arising out of a particular shipment would be barred and those in favor of the other party not be barred.
Pennsylvania R. Co., 16 F.2d at 761.
Although the explanation of “accrues” in § 415(e) is similar to that contained in the corresponding provision of the Commerce Act, those two provisions rest upon totally different policy considerations.
The [Commerce Act] is explicit that accrual under § 16(S)(e) is not synonymous or mutually interchangeable with collectability or dueness of the debt and that the point of delivery overrides any conventional notion of when an action judicially matures _ [C]ongress anticipated the multitude of variations in the period of limitations that would result if conventional standards of triggering the period were used and promulgated 16(3)(a) and (e) in order to avoid such variation by creating a uniform time of accrual irrespective of when the action could have generally been brought.
South Omaha Terminal Ry. Co., 373 F.Supp. at 644. Different considerations arise when the conflict is between a telecommunications company and its customer. It would be nonsensical to conclude that a cause of action to collect charges for transmission of a telephone call accrues when a message is delivered. Such a rule would allow a carrier to sue a customer as soon as the carrier completes a telephone connection even if the customer has not refused to pay, and even if the customer has given every indication that it would pay upon receipt of an accurate bill. When a call is completed, the carrier has typically not even billed for the service, and the customer has every right to assume that it will not incur any liability so long as it pays the bill that is expected at some future time.
However, where a customer seeks redress from a telecommunications carrier based upon the improper transmission of a telecommunications signal, it is reasonable to conclude that the cause of action in favor of the customer does arise when the message is delivered or tendered. It is then that the customer reasonably should know of the alleged problem. See Central Scott Tel. Co. v. Teleconnect Long Distance Servs. & Sys. Co., 882 F.Supp. 1317, 1320-21 (S.D.Iowa 1993)10 (the court relied upon the FCC’s declaration that § 415(e) only “concerns a carrier’s liability to its customers for failure to transmit a message in accordance with its common carrier obligations”). See also Anchorage Tel. Util. v. Alascom, Inc., 4 FCC Red. 2472 (1989) (“Section 415(e) is concerned with a carrier’s liability to its customer for failure to transmit a message ... ”); MCI Telecommunications Corp. v. Pac. Bell Tel. Co., 5 F.C.C. Red. 3462 (1990) (“Section 415(e) is inapplicable to MCI’s complaints. That Section concerns a carrier’s liability to its customers ...”).
We agree that the limitations period contained in § 415(e) applies to actions brought by the carrier’s customer that allege a breach of a common carrier’s obligation and not to an action against the customer. Accordingly, here, the statute of limitations for purposes of § 415(a) accrues with “discovery of the right or wrong or of the facts on which such knowledge is chargeable in law.” Central Scott Tel. Co., 832 F.Supp. at 1320. MCI’s tariff tells us when that occurred. Under that tariff, “MCI’s bills are payable upon receipt,” however, “[a]mounts not paid within 30 days after the date of the invoice will be considered past due.... ” We must give these words their ordinary meaning. See Strite v. McGinnes, 330 F.2d 234, 239 (3d Cir.), cert. denied, 379 U.S. 836, 85 S.Ct. 69, 13 L.Ed.2d 43 (1964) (the words employed are given “their plain and ordinary meaning, except where the context in which they are used renders then a different denotation, or where legal or technical words are used and it is clear from their use that the legal or technical meaning was intended.”). Teleconcepts’ obligations became past due 30 days after the date of a particular invoice. It is therefore only then that MCI’s cause of action under § 415(a) accrued, and it is at *1101that point that the two year clock started ticking.
Moreover, Teleconeepts’ duty (both under the tariff and as reflected by MCI’s actual billing practice) is akin to an obligation to make installment payments. “In an installment contract, a new cause of action arises from the date each payment is missed.” Board of Trustees of the Dist. No. 15 Machinists’ Pension Fund v. Kahle Eng. Corp., 43 F.3d 852, 857 (3d Cir.1994) (citing 4 A. Corbin, Corbin on Contracts § 951 (1951)). “[T]he statute of limitations runs against each installment from the time it becomes due, that is, from the time when an action might be brought to recover it.” Id. at 857 (quoting 51 Am.Jur.2d: Limitations of Actions § 133). See also Metromedia Co. v. Hartz Mountain Assoc., 139 N.J. 532, 655 A.2d 1379, 1380-81 (1995) (contract between lessee and lessor whereby lessor was to reimburse lessee for lessee’s use of an independent cleaning service treated like an installment contract and thus a new cause of action subject to its own limitations period accrued for each month that lessor failed to reimburse the lessee); Kiamichi Electric Cooperative v. Underwood, 842 P.2d 358, 359-60 (Okla.Ct.App.1992) (electric cooperative’s contract to provide electricity was atón to installment contract and thus each monthly installment due constituted a new cause of action with its own statute of limitations).
The period fixed by a statute of limitations begins to run from the ‘accrual of the cause of action.’ Since ‘cause of action’ is so uncertain and variable a concept, serious injustice may be done unless the court-uses judicial discretion in applying such a statute in the case of ‘partial’ breaches of a single contract. No doubt there is much authority for the statement that where separate actions would lie for a series of breaches, the statute operates against each one separately as of the time when each one could have been brought, and that this rule is not affected by the fact that after two or more such breaches have occurred the plaintiff must join them all in one cause of action. Of course, if an action for a first instalment is barred by the statute, it cannot be properly included in an action for later installments that are not yet barred.
Corbin, supra, § 951, at 823-24.
Here, MCI first accrued a cause of action on December 8,1989—30 days after the date of the November 8, 1989 long distance bills. Additional causes of action accrued 30 days after the date of each of MCI’s subsequent bills. MCI could have instituted suit to recover payment on any one bill, or brought an action as they did to recover payment on all of them. However, for MCI’s action to have been timely as to all of the bills it must have been filed within two years of the date on which a cause of action accrued on each partial breach. See Kahle Engineering Corp., 43 F.3d at 861 (action barred as to those unpaid installments which came due prior to six year statute of limitations of the MPPAA); Metromedia, 655 A.2d at 1381; Corbin, supra, § 951, at 823-24.
Thus, MCI’s action is untimely as to the amounts owed in the bills dated November 8, 1989, November 15, 1989, and December 8, 1989. MCI’s cause of action for these bills accrued on December 8, 1989, December 15, 1989 and January 8, 1990, respectively. Since the complaint was not filed until January 15, 1992, recovery for those bills is barred by the two year statute of limitations of § 415(a).
However, MCI’s action is timely as to all subsequent bills. Accordingly, MCI can maintain its suit for amounts owing on the bills dated December 15, 1989;11 January 8, 1990; January 9, 1990; February 8, 1990; March 8, 1990; and March 17, 1990.
Accordingly, we reverse the judgment of the district court and remand for the court to deduct from its judgment the amount of the bills dated November 8, 1989, November 15, 1989, and December 8, 1989 and to recalculate the award of prejudgment interest accordingly.
E. Attorney’s Fees
Teleconeepts challenges the award of attorney’s fees arguing that hours expended by MCI’s counsel were “clearly exces*1102sive,” since there was no discovery and most of the work merely involved the preparation of pleadings. See Appellant’s brief at 32-38.
The district court must exercise its informed discretion in awarding attorney’s fees. See Pawlak v. Greenawalt, 713 F.2d 972, 977 (3d Cir.1983), cert. denied, 464 U.S. 1042, 104 S.Ct. 707, 79 L.Ed.2d 172 (1984). Thus, our standard of review is a narrow one. “We can find an abuse of discretion if no reasonable [person] would adopt the district court’s view. If reasonable [people] could differ as to the propriety of the action taken by the trial court, then it cannot be said that the trial court abused its discretion.” Silberman v. Bogle, 683 F.2d 62, 65 (3d Cir.1982) (citation omitted).
The district court was familiar with the efforts of counsel, and conducted a careful review of the time that counsel spent working on this case. The court found that the “actual hours claimed were in fact reasonably expended by counsel.” MCI Telecommunications Corp. v. Teleconcepts, Inc., No. 92-244, slip op. at 3 (D.N.J. February 25, 1994). Teleeoneepts does not specify how the district court abused its discretion in its review, and we do not think that any such abuse of discretion occurred. Accordingly, we affirm the award of attorney’s fees. However, in view of our ruling on the statute of limitations, the district court should, upon remand, make whatever review of attorney’s fees it feels warranted and adjust the prior award of fees if the court feels that such a reduction or adjustment is now appropriate. We do not, however, take any position as to whether the court should make any such adjustment.
F. Third-Party claim
Finally, Teleeoneepts argues that the district court erred in granting summary judgment in favor of third-party defendant, Bell. Our review of this grant of summary judgment is plenary. See Gulfstream III, 995 F.2d at 429.
Teleeoneepts maintains it is entitled to be indemnified for any liability it owes to MCI because Bell allowed the fraudulent “hacking” to occur by furnishing a defective dial tone. Teleeoneepts’ third-party complaint does not allege the basis for the district court’s subject matter jurisdiction, nor does the district court state the basis of its jurisdiction. However, since there was no independent basis for subject matter jurisdiction over that claim, we believe that the district court exercised supplemental jurisdiction based on the FCC tariff. See 28 U.S.C. § 1367 (Supp.1993).12
Congress codified the judicially created doctrines of pendent and ancillary jurisdiction under the name “Supplemental Jurisdiction” at 28 U.S.C. § 1367. Section 1367 embodies the jurisdictional standard established in United Mine Workers of America v. Gibbs, 383 U.S. 715, 86 S.Ct. 1130, 16 L.Ed.2d 218 (1966). See Lyon v. Whisman, 45 F.3d 758, 760 (3d Cir.1995); Sinclair v. Soniform, Inc., 935 F.2d 599, 603 (3d Cir.1991). Accordingly, three requirements must be satisfied before a federal court may exercise supplemental jurisdiction. “The federal claim must have substance sufficient to confer subject matter jurisdiction on the court.” Gibbs, 383 U.S. at 725, 86 S.Ct. at 1138. The state and federal claims must derive from a common nucleus of operative facts, and the claims must be such that they would ordinarily be expected to be tried in one judicial proceeding. See id.; Lyon, 45 F.3d at 760.
We believe that the third party complaint satisfies the prerequisites of § 1367(a). 28 U.S.C. §§ 1331, and 1337 confer subject matter jurisdiction over MCI’s claim. MCI’s action and Teleeoneepts’ third-party action both arise out of the fraudulent “hacking” activity which purportedly resulted in exorbitant long distance charges. Moreover, logic and prudent use of judicial resources dictate that these claims be tried in one judicial proceeding. Accordingly, the third-party action meets the test for supplemental jurisdiction.
*1103However, our inquiry does not end there. Bell argues that the tariff it filed with the Pennsylvania Public Utilities Commission (“P.U.C.”) places responsibility for unauthorized calls upon Teleconeepts. That tariff provides in pertinent part: “The COCOT13 ... [s]ervice subscriber is considered as the Customer of Record and is responsible for all rates and charges associated with the service,....” Resolution of the third-party complaint therefore turns upon the interpretation and application of this PUC tariff. We must therefore determine if the doctrine of primary jurisdiction applies.
Primary jurisdiction ‘applies where a claim is originally cognizable in the courts, and comes into play whenever enforcement of the claim requires resolution of issues which, under a regulatory scheme, have been placed within the special competence of an administrative body.’ In contrast, when the legislature provides an agency with ‘exclusive primary jurisdiction,’ it preempts the courts’ original jurisdiction over the subject matter.
Greate Bay Hotel & Casino v. Tose, 34 F.3d 1227, 1230 n. 5. (3d Cir.1994) (citations omitted). If a legislature has vested an administrative agency with exclusive primary jurisdiction, that agency is the only forum in which complaints within that jurisdiction may be brought. Id. at 1230. Even though the parties have not raised the doctrine of exclusive primary jurisdiction, we must determine if the third-party complaint should be heard by the PUC in the first instance.14 We are mindful of the fact that the concept of primary jurisdiction:
is not simply a polite gesture of deference to the agency seeking an advisory opinion wherein the court is free to ignore the agency’s determination. Rather, once the court properly refers a matter or a specific issue to the agency, that agency’s determination is binding upon the court and the parties (subject, of course, to appellate review through normal channels), and is not subject to collateral attack in the pending court proceeding.
Elkin v. Bell Tel. Co. of Pa., 491 Pa. 123, 420 A.2d 371, 376 (1980) (footnotes omitted).
1. The Statutory Framework of the PUC
“The PUC has long been recognized as the appropriate forum for the adjudication of issues involving the reasonableness, adequacy and sufficiency of public utility services.” Behrend v. Bell Telephone, 431 Pa. 63, 243 A.2d 346, 347 (1968).
The PUC has the power to ‘prescribe as to service and facilities ... just and reasonable standards ... to be furnished, imposed, observed, and followed by any or all public utilities ...’ and upon finding, ... ‘that the service or facilities of any public utility are unreasonable, unsafe, inadequate, insufficient ... ’ the PUC ‘shall determine and prescribe, by regulation or order, the reasonable, safe, adequate, sufficient, service or facilities to be observed, furnished, enforced or employed ...’
Elkin v. Bell Telephone, 491 Pa. 123, 420 A.2d 371, 374 (1980), see 66 P.S. §§ 1182, 1183, (1959) (repealed and replaced by 66 Pa.C.S. §§ 1504, 1505 (1978)). The Pennsylvania Public Utility Law requires public utilities to file tariffs with the PUC. See 66 Pa.C.S. § 1302 (Purdon 1979 & Supp.1995). These tariffs are binding and dispositive of the rights and liabilities between the customer and the public utility. See 66 Pa.C.S. § 1303 (Purdon 1979). The PUC has enforcement power over its tariffs and regulations, and matters that pertain to those tariffs are considered to be within the particular expertise of the PUC. See 66 Pa.C.S. § 501, et seq. (Purdon 1979).
Accordingly, “[t]he PUC has long been recognized as the appropriate forum for the adjudication of issues involving the reasonableness, adequacy and sufficiency of public utility sendees.” Elkin, 420 A.2d at 374. At oral argument, Teleconcepts conceded that it was challenging the reasonableness, adequacy and sufficiency of Bell’s telephone service, and Bell admitted this point in its *1104supplemental brief to this court. See Bell's Supp. Brief at 1-2, 7. Thus, it is not disputed that the subject matter of the third-party complaint is within the jurisdiction of the PUC, and that agency is the appropriate forum to resolve the issues raised by that complaint. That determination, however, though necessary to our analysis, is not sufficient to end our inquiry. Our inquiry must then focus upon whether resolution of Tele-concepts’ claim against Bell requires the special competence of the PUC.
Courts should not be too hasty in referring a matter to an agency, or to develop a ‘dependence’ on the agencies whenever a controversy remotely involves some issue falling arguably within the domain of the agency’s ‘expertise.’ ‘Expertise’ is no talisman dissolving a court’s jurisdiction. Accommodation of the judicial and administrative functions does not mean abdication of judicial responsibility....
Therefore, where the subject matter is within an agency’s jurisdiction and where it is a complex matter requiring special competence, with which the judge or jury would not or could not be familiar, the proper procedure is for the court to refer the matter to the appropriate agency.... Where, on the other hand, the matter is not one peculiarly within the agency’s area of expertise, but is one which the courts or jury are equally well-suited to determine, the court must not abdicate its responsibility.
Elkin, 420 A.2d at 377.
2. The Need for the PUC’s Expertise
As noted above, issues that implicate a utility’s tariff are deemed to be within the special expertise of the PUC. In addition we are guided by Elkin, and DeFrancesco v. Western Pa. Water Co., 499 Pa. 374, 453 A.2d 595, 596 (1982). In Elkin, the court held that allegations that Bell negligently failed to furnish the customer “reasonable rapid and efficient service” with respect to three wide-area telephone service (“WATS”) lines, deliberately refused to furnish plaintiff with adequate directory assistance information service, and negligently failed to furnish written telephone numbers for prospective customers of plaintiff fell within the PUC’s area of expertise. Elkin, 420 A.2d. at 373, 377.
By contrast, in DeFrancesco the court held that the allegation that fire damaged plaintiffs property because the city water company negligently failed to maintain proper water pressure did not fall within the PUC’s expertise and thus its primary jurisdiction. The court reasoned:
The controversy now before us ... is not one in which the general reasonableness, adequacy or sufficiency of a public utility’s service is drawn into question. Resolution of appellant’s claims depended upon no rule or regulation predicated on the peculiar expertise of the PUC, no agency policy, no question of service or facilities owed the general public, and no particular standard of safety or convenience articulated by the PUC_ Rather,_ [r]esolving the essential question of whether the utility failed to perform its mandated duties requires no recondite knowledge or experience and falls within the scope of the ordinary business of our courts.
DeFrancesco, 453 A.2d at 597.
Courts have routinely looked to Elkin and DeFrancesco to determine if a particular controversy implicated the special competence of the PUC. See Optimum Image, Inc. v. Phila. Elec. Co., 410 Pa.Super. 475, 600 A.2d 553, 556-57 (1991) (allegations that tariff was violated by the substandard and defective supply of electrical power brought controversy within the primary jurisdiction of the PUC); Ostrov v. I.F.T., Inc., 402 Pa.Super. 87, 586 A.2d 409, 415 (1991) (action did not come within the primary jurisdiction of PUC since plaintiff did not contend that medical examination provision of self-insurance plan violated the PUC’s rules or regulations governing self-insurance motor carriers nor any other PUC rule or regulation for self insurance) Schriner v. Pa. Power & Light Co., 348 Pa.Super. 177, 501 A.2d 1128, 1130-31 (1985) (citing DeFrancesco the court held that problem of “stray voltage” depends upon “no rule or regulation predicated upon the peculiar expertise of the PUC ...” and thus was not within the primary jurisdiction of the PUC); Morrow v. Bell Tel. Co. of Pa., 330 Pa.Super. 276, 479 A.2d 548, 551-52 (1984) (plaintiffs *1105challenge to public utility’s rates relating to toll charges and its service practice regarding deposits held to be within the primary jurisdiction of the PUC).
Here, the dispute centers around Bell’s performance under its tariff and any technical deficiencies that may have existed in the dial tone generated by its equipment. That complaint may rise or fall on the issue of the manner in which Bell complied with its obligation under its tariff to provide “reasonable, ... efficient” service. The agency that can best determine Bell's compliance with that tariff is the PUC. In addition, Teleeoneepts’ allegation of deficient service transcends the present controversy and, at least potentially, calls into question the adequacy of Bell’s service to the general public as Teleeoneepts claims that the second dial tone was neither unique to Teleeoneepts nor COCOT owners. We must therefore be sensitive to the need for uniformity and consistency in agency policy, which further suggests that the PUC decide the merits of Teleeoneepts’ claim initially. See Elkin, 420 A.2d at 377; Ostrov, 586 A.2d at 415 (“matters involving the general reasonableness or adequacy of a utility’s service to the public are within the primary jurisdiction of the [PUC].”).
3. The Effect of the Relief Sought
Nevertheless, despite the need for the PUC to determine liability on Teleeoneepts’ third-party claim, the PUC is not empowered to award damages, and Teleeoneepts is seeking damages in the nature of indemnification. See Elkin, 420 A.2d at 374 (“the PUC has no authority to award damages”). However, the Pennsylvania Supreme Court has approved a bifurcated procedure where, as here, a plaintiff sues a public utility based upon the latter’s purported failure to provide adequate, reasonable or sufficient service, but seeks damages as a remedy. See Elkin, 420 A.2d at 375-76; Ostrov, 586 A.2d at 414. Under this bifurcated procedure, the issue of liability is transferred to, and initially decided by, the PUC. If necessary, the appropriate trial court thereafter determines damages. See Elkin, 420 A.2d at 377; DeFrancesco, 453 A.2d at 596 n. 3.
Accordingly, “the doctrine of primary jurisdiction applies where the administrative agency cannot provide a means of complete redress to the complaining party and yet the dispute involves issues that are clearly better resolved in the first instance by the administrative agency charged with regulating the subject matter of the dispute.” Ostrov, 586 A.2d at 413. The doctrine requires a court to transfer an issue that involves administrative expertise to the administrative agency charged with exercising that discretion. Richman Bros., 953 F.2d at 1435 n. 3. “Essentially, the doctrine creates a workable relationship between the courts and administrative agencies wherein, in appropriate circumstances, the courts can have the benefit of the agency’s views on issues within the agency’s competence.” Elkin, 420 A.2d at 376.
In Optimum Image, supra, the plaintiff sued the Philadelphia Electric Company (“PECO”) alleging that the utility “wrongfully, negligently, carelessly and without reasonable cause delivered, over an extended period of time, unreasonably defective electrical power” to plaintiffs business premises. The trial court transferred the determination of liability to the PUC. The Pennsylvania Superior Court affirmed.
[Plaintiff alleges that] the electrical power supplied by PECO exceeded the ten percent variation allowed by PECO’s tariff filed with the PUC. In addition, [plaintiff] ... alleges that the power with which it was supplied was substandard and outside the regulatory requirements and that the problem it experienced was not investigated with the proper equipment.
In response, PECO contends that it at all times provided electrical power in compliance with its tariff filed with the PUC and otherwise furnished and maintained adequate, efficient, safe and reasonable services and facilities to [plaintiff].
Id. 600 A.2d at 556-57. The controversy between Teleeoneepts and Bell is analogous. We believe that the doctrine of primary jurisdiction required the district court to utilize the bifurcated procedure established for resolving questions of liability where damages *1106are sought in a matter involving the special expertise of the PUC. Thus, although the district court had jurisdiction over the third-party claim, the court erred in deciding the question of liability. That claim must be transferred to the PUC for such a determination. If the PUC concludes that Telecon-cepts is entitled to indemnification from Bell, the district court may then make an appropriate award to Teleconcepts. See Optimum Image, 600 A.2d at 557.
We are aware that language in some decisions of the United States Supreme Court and our sister Courts of Appeals seems to suggest a contrary result here. For example, in Reiter v. Cooper, 507 U.S. 258, 113 S.Ct. 1213, 122 L.Ed.2d 604 (1993) the Court stated:
Referral of the issue to the administrative agency does not deprive the court of jurisdiction; it has discretion to either retain jurisdiction or, if the parties would not be unfairly disadvantaged, to dismiss the case without prejudice.
507 U.S. at -, 113 S.Ct. at 1219. See also, U.S. v. Philadelphia National Bank, 374 U.S. 321, 83 S.Ct. 1715, 10 L.Ed.2d 915 (1963) (comparing primary jurisdiction to a prudential doctrine of abstention and noting that primary jurisdiction merely postpones and does not preclude the exercise of jurisdiction by a federal court), Northwest Airlines, Inc. v. County of Kent, Mich., — U.S. -, 114 S.Ct. 855, 127 L.Ed.2d 183 (1994) (failure to brief primary jurisdiction resulted in waiving consideration of the doctrine), Gross v. Baxter Healthcare Corp., 51 F.3d 703, 706 (7th Cir.1995) (the court found the parties had waived any consideration of primary jurisdiction and noted, “[i]n this respect, primary jurisdiction is quite different from subject matter jurisdiction[ ]”), and U.S. v. Henri, 828 F.2d 526, 527 (9th Cir.1987) (“the doctrine of primary jurisdiction, despite what the term may imply, does not speak of the jurisdictional power of the federal courts[]”). However, none of these eases addressed the issue of the authority of a federal court to adjudicate a matter that a state legislature had placed within the exclusive domain of a state administrative agency. Accordingly, our analysis here is consistent with the results reached in such cases. See U.S. v. Western Pacific Railroad, 352 U.S. 59, 63, 77 S.Ct. 161, 165, 1 L.Ed.2d 126 (1956). There, the court stated:
[t]he doctrine of primary jurisdiction thus does ‘more than prescribe the mere procedural timetable of the law suit. It is a doctrine allocating the law making power over certain aspects’ of commercial relations. ‘It transfers from court to agency the power to determine’ some of the incidents of such relations.
352 U.S. at 65, 77 S.Ct. at 166 (emphasis added). See also, Slocum v. Delaware, L. & W.R. Co., 339 U.S. 239, 70 S.Ct. 577, 94 L.Ed. 795 (1950), and Pennsylvania Railroad Co. v. Day, 360 U.S. 548, 79 S.Ct. 1322, 3 L.Ed.2d 1422 (1959), and Greate Bay Hotel and Casino, supra.
The Commonwealth of Pennsylvania has committed the issues raised by Teleconcepts’ claim against Bell to the exclusive jurisdiction of the Pennsylvania Public Utilities Commission and a federal court cannot amend state law by exercising supplemental jurisdiction. Indeed, a contrary holding would mean that the federal courts are empowered to decide matters of state law that courts in the affected state lack authority to resolve. Since the Supreme Court of Pennsylvania has held that “the Public Utility Commission has been vested by the legislature with exclusive original jurisdiction ...” of the issues which this suit requires us to resolve, Behrend v. Bell Telephone Co., 431 Pa. 63, 243 A.2d 346, 347 (1968), we must remand to the district court for appropriate proceedings.
Accordingly, we will reverse the district court’s grant of summary judgment in favor of Bell and remand so that the district court can transfer the third-party claim to the PUC for a determination of liability. The district court retains jurisdiction over this matter pending the liability determination by the PUC.15
*1107III.
For the reasons stated above, we will affirm in part and reverse in part the judgments of the district court, and remand this case to the district court for proceedings consistent with this opinion.
. It is unclear exactly when Teleeoncepts first learned that it may have been a "hacking" victim. In an affidavit filed in the district court, John Goida, president of Teleeoncepts, states he learned generally of the fraudulent "hacking" process during an unrelated phone conversation with an employee in Bell's fraud division, and he deduced that this is what happened to his pay phones.
. It appears that the defect in the dial tone has since been remedied.
. 28 U.S.C. § 1331 provides that "the district courts shall have original jurisdiction of all civil actions arising under the Constitution, laws, or treaties of the United States.” 28 U.S.C. § 1337(a) provides, in part, that "[t]he district courts shall have original jurisdiction of any civil action or proceeding arising under any Act of Congress regulating commerce or protecting trade and against restraints and monopolies. ...”
. Although some courts that have addressed this issue have relied upon § 1331 and others have relied upon § 1337, there is no difference in these two bases of subject matter jurisdiction for purposes of our present analysis. See Medlin, 620 F.2d at 962-963 ("The ‘arising under' requirement of section 1337 has been interpreted to be the same as that found in 28 U.S.C. § 1331....”); Yancoskie v. Delaware River Port Authority, 528 F.2d 722, 725 (3d Cir.1975) (same).
. In Thurston, the Court rejected the Ninth Circuit Court of Appeals’ argument that a carrier's action for payment of transportation services was a “simple contract collection action." Garden State Investment, supra, at 387 (citing Thurston, 460 U.S. at 533, 103 S.Ct. at 1343).
. As of December 1, 1993, Rule 4(j) was amended and redesignated Rule 4(m). We discuss the significance of this amendment infra.
. Fed.R.Civ.P. 6(b) provides in pertinent part: Enlargement. When by these rules or by a notice given thereunder or by order of court an act is required or allowed to be done at or within a specified time, the court for cause shown may at any time in its discretion (1) with or without motion or notice order the period enlarged if request therefor is made before the expiration of the period originally prescribed or as extended by a previous order, or (2) upon motion made after the expiration of the specified period permit the act to be done where the failure to act was the result of excusable neglect....
. In 1978, after the cases relied upon by Telecon-cepts had been decided, 49 U.S.C. § 16(3) was repealed and replaced with 49 U.S.C. § 11706.
. "Accrual” is now defined in 49 U.S.C. § 11706(g) which states "[a] claim related to a shipment of property accrues under this section on delivery or tender of delivery by the carrier.” 49 U.S.C. § 11706 (1982).
. Citing Williams Telecommunications Group, Inc. v. The Chesapeake & Potomac Tel. Cos., 8 FCC Red. 1161, ¶ 16 n. 35 (1993).
. Suit was instituted two years to the day from the date the cause of action accrued on this bill.
. Diversity jurisdiction is clearly inapplicable since the amount in controversy does not exceed $50,000. See 28 U.S.C. § 1332 (1988).
. "COCOT” is an acronym for "Customer-Owned Coin-Operated Telephones.” All parties agree that Teleconcepts is a COCOT within the meaning of that term in the PUC tariff.
. Following oral argument we asked the parties to file supplemental briefs on this issue.
. We realize that after the district court disposes of the allocation of long distance charges between MCI and Teleconcepts in accordance with this opinion, the court will be left with only the third-party claim founded on state law. It is for the district court to decide whether to retain *1107supplemental jurisdiction at that point, or to dismiss the third-party claim pursuant to 28 U.S.C. § 1367(c)(3).